See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/275043256

Evolution of clay fabric and water retention

properties along hydromechanical stress paths

Conference Paper · June 2014

DOI: 10.1201/b17017-173

CITATIONS READS

2 50

5 authors, including:

Anne-Catherine Dieudonné Severine Levasseur

Delft University of Technology ONDRAF/NIRAS
16 PUBLICATIONS 51 CITATIONS 52 PUBLICATIONS 407 CITATIONS

SEE PROFILE SEE PROFILE

Gabriele Della Vecchia Cristina Jommi

Politecnico di Milano Delft University of Technology
44 PUBLICATIONS 322 CITATIONS 80 PUBLICATIONS 542 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Time dependent response of dykaes and levees View project

Ballasted tracks View project

All content following this page was uploaded by Anne-Catherine Dieudonné on 17 April 2015.

The user has requested enhancement of the downloaded file.

Evolution of clay fabric and water retention properties along
hydromechanical stress paths

A.C. Dieudonne
ArGEnCo Department, University of Liege, Liege, Belgium
F.R.I.A., Fond de la Recherche Scientifique - FNRS, Brussels, Belgium

R. Charlier, S. Levasseur
ArGEnCo Department, University of Liege, Liege, Belgium

G. Della Vecchia
Department of Civil and Environmental Engineering, Politecnico di Milano, Milano, Italy

C. Jommi
Department of Geosciences and Engineering, TU Delft, Delft, The Netherlands

ABSTRACT: The paper presents a water retention model for compacted clayey soils. The model is based
on a description of the aggregated structure of the material and its evolution along generalized hydraulic and
mechanical stress paths. The water retention properties of each structural level are distinguished and described
separately using an expression of the type proposed by van Genuchten (1980). The water retention model is val-
idated against experimental data on Boom Clay compacted at different dry densities. Its qualitative performance
are highlighted along a complex hydromechanical path. The proposed formulation succeeds in tracking simul-
taneously the evolution of the fabric pattern and of the hydraulic state of compacted clays along generalized
stress paths.

1 INTRODUCTION
Compacted clays are used in a wide range of geotech-
nical and geo-environmental applications. In civil en-
gineering, compacted clayey soils are found in em-
bankments and earth dams, whereas they are consid-
ered as engineered barriers for the disposal of domes-
tic, industrial and hazardous waste. In these applica-
tions, the compacted material remains essentially un-
saturated over its service life. The characterization of
its water retention properties is thus fundamental to
a comprehensive understanding and modelling of the
coupled response of the material in terms of water
flow and deformation processes.
Since many years, numerous studies have recog-
nised the importance of the material structure on the
engineered properties of compacted clayey soils (see
Romero 2013 for a review). While compacted on
the dry side of optimum, clayey soils exhibit a typ-
ical aggregated fabric. This fabric is not unique and
the size of the aggregates appears to be sensitive
to the compaction water content (Thom et al. 2007,
Birle 2012), the chemical composition of the inter-
stitial fluid (Musso et al. 2013) and the particular hy-
dromechanical path followed during compaction (De-
lage 2010). In addition, the structure is not fixed but
evolves along with mechanical, hydraulic or chem-
ical loading (see for instance Romero et al. 1999,
Cuisinier and Laloui 2004, Della Vecchia 2009, Mon-
roy et al. 2010, Romero et al. 2011, Casini et al. 2012,
Wang et al. 2013).
In the last years, increasing attention has been
paid to the links between the water retention prop-
erties of a soil and its evolving pore size distribu-
tion (PSD). Durner (1994) modelled successfully the
water retention behaviour of a sandy loam, distin-
guishing explicitly in the formulation the existence
of two structural levels. The same approach was later
used by Gitirana Jr. and Fredlund (2004) on a pel-
letized diatomaceous soil and on a residual, highly
collapsible clay from Brasilia. Romero and Vaunat
(2000) distinguished an intra-aggregate water region
and inter-aggregate water region to model the reten-
tion behaviour of Boom Clay. However none of these
Authors have considered explicitly the evolutionary
 0.8
character of the material structure along hydrome-

Pore size density function, -de/d(log x)

Macrostructural level
0.7
chanical stress paths. Early attempts to include the (inter-aggregate porosity)

evolution of the structure into a water retention model 0.6

are due to Simms and Yanful (2002, 2004), Romero 0.5

et al. (2011) and Della Vecchia et al. (2013). 0.4

Microstructural level
In this paper, a water retention model for com- 0.3 (intra-aggregate porosity)

pacted clayey clays is proposed. The model takes 0.2

into account the aggregated structure of the material 0.1
as it is directly linked to its pore size density func-
0 0
tion through Laplace equation. The model is validated 10 101 102 103 104 105
Entrance pore size, x [nm]
against experimental data on compacted Boom Clay.
Figure 1: Theoretical pore size distribution derived
Finally the model capabilities are assessed along a
from the proposed water retention model.
complex hydromechanical stress path.
can be derived directly from equation (1):
2 THEORETICAL FORMULATION
∂ew 4σ cos θw ∂ew
P SD ≡ =− (3)
The water retention model is written as the superpo- ∂ log x x log e ∂s
sition of two elementary curves, using the approach
proposed by Othmer et al. (1991) and Durner (1994). where σ = 0.07275 N/m is the air/water surface ten-
Each elementary curve is used to describe the water sion and θw = 0◦ is the water-solid contact angle.
retention properties of a structural level. In this pa- Figure 1 presents the theoretical pore size distri-
per, van Genuchten model (van Genuchten 1980) is bution derived from the water retention model. The
adopted for the description of the retention behaviour model parameters s0 , n and m can thus be interpreted
of both microstructural and macrostructural levels. with reference to the pore size density function. The
The water ratio ew , defined as the volume of water value of s0 is related to the position, on the x-axis,
over the volume of solids, is given by: of the maximum values of the PSD function, while n
and m are associated to the width and the shape of the
pore size distribution.
 !n(m) −m(m) Dieudonne et al. (2013) have highlighted the evo-
s lution of the microstructural and macrostructural pa-
ew (s) = em 1 + (m)
 (m) (M )
s0 rameters s0 and s0 along hydraulic and mechan-
ical loading stress path. Experimental PSD data on
compacted Boom Clay from Della Vecchia (2009)
 !n(M ) −m(M ) and on compacted London Clay from Monroy et al.
s (2010) were used to calibrate the model. According
+ (e − em ) 1 + (M )
 (1)
(m)
s0 to experimental evidence, the parameters s0 and
(M )
s0 are assumed to evolve exponentially respectively
where s is the suction, e and em are the total and mi- with the microstructural void ratio em and with the
(m) (M )
crostructural void ratios, s0 and s0 are measures ratio (e − em )/e, representing the ratio between the
of the microstructural and macrostructural air-entry macrostructural and the total void ratios:
pressures and n(m) , m(m) , n(M ) and m(M ) are model (m) (m)

(m)

parameters. s0 = α1 exp −α2 em (4)
The dependency of the microstructural void ra-  
tio em on the water content has been highlighted by (M ) (M ) (M ) e − em
(Romero et al. 2011). In this paper, the following re- s0 = α1 exp −α2 (5)
e
lationship is used:
where α1 and α2 are model parameters (Figure 2).
em = β0 e2w + β1 ew + em,0 (2) Equations (1) - (5) describe a state surface in the
suction - void ratio - degree of saturation space. This
where β0 and β1 quantify the swelling tendency of surface is depicted in Figure 3 for Boom Clay, using
the aggregates and em,0 is the microstructural void ra- the model parameters presented in Table 1. The de-
tio of the dry material. This equation is similar to the pendency of the water retention curve on the void ra-
one proposed by Romero et al. (2011), except that it tio is clearly highlighted.
suggests a continuous evolution of the microstructural
void ratio with water ratio. 3 EXPERIMENTAL VALIDATION
Considering the Washburn equation (Washburn
1921), relating the suction s to an equivalent pore di- The proposed water retention model has been vali-
ameter x, a theoretical pore size density (PSD) model dated against experimental data on Compacted Boom
 2
10

Macrostructural air-entry value, s0 [MPa]

2
Microstructural air-entry value, s0 [MPa] 10
Boom Clay

(M)
Calibrated values
(m)

0
Least-squares exponential fitting 10
1 s(m)
0 = 180 exp(-9 em) [MPa]
10
R2 = 0.89

Boom Clay
-2
10
0 Calibrated values
10
Least-squares exponential fitting
s(M)
0 = 1.5 exp(-5 eM/e) [MPa]
R2 = 0.75
-4
10
-1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
10
0.2 0.3 0.4 0.5 0.6 0.7 0.8 Macrostructural void ratio over total void ratio, eM/e [-]
Microstructural void ratio, em [-]
(b) Dependence of the macrostructural air-entry pressure
(a) Dependence of the microstructural air-entry pressure with the ratio of the macrostructural void ratio over the total
with the microstructural void ratio. void ratio.
Figure 2: Variation of the model parameters along hydromechanical stress paths.

1
Clay from Romero et al. (2011). The model parame- Experimental data
(Romero et al., 2011)
0.9
ters were first determined by calibrating the pore size Axis translation (e = 0.93)
0.8 Axis translation (e = 0.65)
distribution model against experimental PSD curves Vapour equilibrium (e = 0.93)

from Della Vecchia (2009). In total, 18 experimental Water ratio, ew [-] 0.7 Vapour equilibrium (e = 0.65)
SMI psychrometer (e = 0.5 - 0.82)
PSD curves were fitted with equation (3) in a system- 0.6 WPA psychrometer (e = 0.93 - 0.99)

atic way in order to highlight the influence of various 0.5 Model (e = 0.93)
Model (e = 0.65)
hydromechanical stress paths, such as wetting, drying 0.4
and loading. Details concerning the calibration pro- 0.3
cess have been presented in Dieudonne et al. (2013). 0.2
In accordance with the calibration of the pore size 0.1
distribution model, the parameters n(m) , m(m) , n(M ) 0 -2 -1 0 1 2 3
and m(M ) are set constant. All parameter values are 10 10 10 10
Suction, s [MPa]
10 10

presented in Table 1. Regarding the evolution of the Figure 4: Comparison between experimental main
microstructural void ratio with the water content, the drying paths for compacted Boom Clay (Romero et al.
following parameters are used: β0 = 0.2, β1 = 0.05 2011) and model predictions at different void ratios.
and em,0 = 0.33.
The performance of the model is presented in Fig- the contrary, for low values of suction, the water re-
ure 4. It can be observed that the model predictions tention curve is sensitive to mechanical actions. Be-
compare favourably with the experimental data on ing calibrated on a huge number of hydromechanical
Boom clay compacted at different void ratios (e = paths, the model can be fruitfully used to predict the
0.65 and e = 0.93). The model succeeds in tracking hydraulic response of the material along different hy-
the increase of air entry pressure with decreasing void dromechanical paths.
ratio. Moreover, for high suction values, the model
tends to reach a unique relationship between water ra-
tio and suction, regardless the current value of void 4 ILLUSTRATIVE EXAMPLE
ratio. In this domain, the water retention is indeed
dominated by the behaviour of the microstructure. On The qualitative response of the water retention model
is evaluated along a complex loading wetting unload-
ing path. The mechanical behaviour of the material is
modelled using the Barcelona Basic Model (Alonso
et al. 1990) so that the total volumetric strain includes
two distinct contributions: a first contribution related
Table 1: Parameters of the water retention model for
Boom Clay.
Microstructural parameters Macrostructural parameters
n(m) 1.65 n(M ) 2.0
m(m) 0.35 m(M ) 0.16
(m) (M )
α1 180 MPa α1 1.5 MPa
(m) (M )
Figure 3: Predicted state surface in the (s-e-Sr ) space. α2 9.0 α2 5.0
to changes in the mean net stress pressure, and a sec- sion of the type proposed by van Genuchten (1980).
ond one associated to changes in suction. Using Washburn equation, a pore size density model
The parameters of Boom Clay have been used in is derived. Selected experimental data of pore size
the simulations. An initial void ratio of 0.97 and an distribution on compacted clays were analysed and
initial suction of 3 MPa are considered, leading to an reinterpreted in the eyes of the proposed formulation
initial degree of saturation equal to 0.55. An aggre- in order to calibrate the model. The water retention
gated structure is associated to this initial state of the model is then validated against experimental data on
material. It is depicted in Figure 5. the same materials compacted at different dry densi-
During a first stage (A), the material is loaded un- ties. The performances of the model are then high-
der oedometer conditions and constant water content lighted along a complex hydromechanical stress path
up to a vertical net stress of 2.5 MPa. During loading, along which the evolutions of the structure and of the
the material experiences a decrease in void ratio (Fig- water retention properties are analysed.
ure 6b). Figure 5 presents the evolution of the pore The proposed model captures important features of
size distribution of the material and shows that load- the retention behaviour of compacted clayey soils,
ing affects mainly the macrostructural pore volume such as the increase in microstructural porous vol-
while maintaining almost constant the volume of mi- ume with increasing water content, the increase in air-
cropores. In addition, the decrease of the porous space entry pressure for decreasing macrostructural voids
is associated with an increase in the degree of satura- and the existence, in the high suction range, of an
tion which is highlighted in Figure 6a. We observe intra-aggregate water region which is almost not sen-
that the reduction in void ratio causes a shifting of the sitive to the total void ratio.
water retention curve towards higher values of suction
for a given degree of saturation.
During a second loading stage, the material is satu- ACKNOWLEDGEMENTS
rated under constant vertical stress (B). The decrease
in suction causes the material to swell, as shown in The authors acknowledge the Belgian FRS-FNRS for
Figure 6b. The saturation process induces an increase its financial support during the stay of the first Author
in the microstructural void ratio. Being under a rela- at the Politecnico di Milano.
tively high constant vertical stress, it is also associated
with a decrease in the macropores diameter (Figure REFERENCES
5). Consequently, the material structure changes from
an aggregated double structure to a monomodal pore Alonso, E., A. Gens, & A. Josa (1990). A constitutive model for
size distribution. partially saturated soils. Geotechnique 40, 405 – 430.
Finally, the material is unloaded under constant Birle, E. (2012). Effect of initial water content and dry density on
water content (C). The unloading process leads to the pore structure and the soil-water retention curve of com-
pacted clay. In C. Mancuso, C. Jommi, and F. D’Onza (Eds.),
swelling of the material which affects mainly the Unsaturated Soils: Research and Applications, pp. 145–152.
macrostructural voids. Casini, F., J. Vaunat, E. Romero, & A. Desider (2012). Conse-
quences on water retention properties of double-porosity fea-
tures in a compacted silt. Acta Geotechnica 7, 139–150.
5 CONCLUSIONS Cuisinier, O. & L. Laloui (2004). Fabric evolution during hy-
dromechanical loading of a compacted silt. Int. J. Num. Anal.
The paper presents a water retention model for Meth. Geomech. 28, 483–499.
compacted clayey soils, accounting for the evolu- Delage, P. (2010). A microstructure approach to the sensitivity
and compressibility of some eastern canada sensitive clays.
tion of their aggregated structure along generalized Geotechnique 60, 353–368.
hydromechanical stress paths. Microstructural and Della Vecchia, G. (2009). Coupled hydro-mechanical behaviour
macrostructural water retention mechanisms are dis- of compacted clayey soils. Ph. D. thesis, Politecnico di Mi-
tinguished and described separately using an expres- lano.
Della Vecchia, G., C. Jommi, & E. Romero (2013). A fully cou-
0.5 pled elastic-plastic hydromechanical model for compacted
Pore size density function, de/d(log x)

soils accounting for clay activity. Int. J. Num. Anal. Meth.

0.4
Geomech. 37, 503–535.
Dieudonne, A., S. Levasseur, R. Charlier, G. Della Vecchia, &
4 Unloaded under C. Jommi (2013). A water retention model for compacted
0.3
constant w clayey soils. In S. Pietruszczack and G. Pande (Eds.), Com-
putational Geomechanics, pp. 23–31.
0.2 Loaded under 2 Durner, W. (1994). Hydraulic conductivity estimation for soils
constant w 1 Initital state
with heterogeneous pore structure. Water Resources Re-
0.1
Saturated under 3
search 30, 211–223.
constant sv Gitirana Jr., G. & D. Fredlund (2004). Soil-water characteristic
0 0 curve equation with independent properties. J. of Geotech.
10 101 102 103 104 105
Entrance pore size, x [nm] and Geoenv. Engng., ASCE 130, 209–212.
Figure 5: Evolution of the pore size distribution dur- Monroy, R., L. Zdravkovic, & A. Ridley (2010). Evolution of
microstructure in compacted London clay during wetting and
ing loading. Geotechnique 60, 105–119.
 1.0
1.00
C
Degree of saturation, Sr [-] Unloading under 0.98
0.9
constant w B Saturation under 0.96

Void ratio, e [-]

constant sv Unloading under
0.8
0.94 constant w
WRC (e = 0.97) WRC Loading under A
(e = 0.885) C constant w
0.7 0.92

0.90 Saturation under B

0.6 constant sv
Loading under constant w A 0.88

0.5 0.86
10-2 10-1 100 101 10-2 10-1 100 101
Suction, s [MPa] Suction, s [MPa]
(a) Stress path in the (s, Sr ) plane (b) Stress path in the (s, e) plane
Figure 6: Evolution of the degree of saturation and the void ratio during a loading - saturation - unloading stress
path performed under oedometer conditions.

Musso, G., E. Romero, & G. Della Vecchia (2013). Double-

structure effects on the chemo-hydro-mechanical behaviour
of a compacted active clay. Geotechnique 63, 206–220.
Othmer, H., B. Diekkruger, & M. Kutilek (1991). Bi-
modal porosity and unsaturated hydraulic conductivity. Soil
Sci. 152, 139–150.
Romero, E. (2013). A microstructural insight into compacted
clayey soils and their hydraulic properties. Engineering Ge-
ology 165, 3–19.
Romero, E., G. Della Vecchia, & C. Jommi (2011). An insight
into the water retention properties of compacted clayey soils.
Geotechnique 61, 313–328.
Romero, E., A. Gens, & A. Lloret (1999). Water permeability,
water retention and microstructure of unsaturated compacted
boom clay. Engineering Geology 54, 117–127.
Romero, E. & J. Vaunat (2000). Retention curves for deformable
clays. In C. Tarantino, A. & Mancuso (Ed.), Experimental
Evidence and Theoretical Approaches in Unsaturated Soils,
Balkema, Rotterdam, pp. 91–106.
Simms, P. & E. Yanful (2002). Predicting soil-water characteris-
tic curves of compacted plastic soils from measured pore-size
distributions. Geotechnique 52, 269–278.
Simms, P. & E. Yanful (2004). Estimation of soil-water charac-
teristic curve of clayey till using measured pore-size distri-
bution. J. Environ. Engng. 130, 847–854.
Thom, R., R. Sivakumar, V. Sivakumar, E. Murray, & P. Mackin-
non (2007). Pore size distribution of unsaturated compacted
kaolin: the initial states and final states following saturation.
Geotechnique 57, 469–474.
van Genuchten, M. (1980). A closed-form equation for predict-
ing the hydraulic conductivity of unsaturated soils. Soil Sci.
Soc. Am. 44, 892–898.
Wang, Q., Y. Cui, A. Tang, J. Barnichon, S. Saba, & W. Ye
(2013). Hydraulic conductivity and microstructure changes
of compacted bentonite/sand mixture during hydration. En-
gineering Geology 164, 67–76.
Washburn, E. (1921). A method of determining the distribution
of pore sizes in a porous material. Proc. Nat. Acad. Sci. 7,
115.

View publication stats